Tracing the Synapses Behind Symphonies:
What songbird neurons can tell us about the brain

In the early hours of dawn in the Santa Cruz Mountains, a conversation is brewing, but not among people. The soft footsteps and whispers of hikers give way to a symphony of tweets and warbles. In one corner of the forest floor, a dark-eyed junco hops from leaf to leaf. Its song is fast—a high-pitched trill of up to 23 notes that fill the space of a few short seconds. Ti-ti-ti-ti-ti-ti-ti-ti.

“It’s another dimension,” says Dave Wilde, gesturing to the chorus above. The park volunteer’s scruffy beard and calm demeanor are exactly what you’d expect from a birder. The 70-year-old is leading a walk through the greens and deep maroons of winter in Henry Cowell Redwoods State Park. As he speaks, the song of a Bewick’s wren rings out behind him—it’s a high-pitched wheeze followed by a swarm of short warbles. Eeee-wee-teu-wewewewewewewewe.

For most in Santa Cruz, this concert of wrens, juncos, and phoebes is quotidian. “People don’t listen. There’s people that have their headphones in,” says Wilde, “What’s that all about?” But this beautiful array of trills and chirps is more than just background noise for neuroscientists, who say it’s a valuable opportunity to better understand how animals have evolved the ability to speak and sing.

Tracing Song

Over the hill, in a lab among redwoods that mirror those of Henry Cowell, UCSC neuroscientist Bradley Colquitt monitors the cage of a Bengalese Finch.

The bird is around 4 inches long and speckled in plumage of white and black.“Sometimes they’ve got funny mop tops,” he jokes, pointing out feathers on the bird’s head reminiscent of a bowl cut. It hops and flutters from perch to perch, letting out soft whirs and mechanical buzzes. Kyut-kuyt-pi-pi-pi-pa. Above it, a small microphone is recording each call.

Hidden within these songs is evidence of the brain’s ability to problem-solve, says Colquitt. “There are very few species that actually learn their vocalizations.” The ability to learn to speak and sing by listening to others—called vocal learning—exists sparingly in the animal kingdom. It’s been found only in cetaceans, bats, elephants, birds, and primates.

Songbird brains are particularly useful for studying vocal learning because of the unique opportunities they provide to isolate the associated neural circuits. “It’s a really fantastic neuroscience model to get a good understanding of how the brain can take auditory information it’s hearing and translate that to control a vocal apparatus,” he says.

Colquitt and his team are studying the brains of these Bengalese finches to get to the bottom of how animals evolved this rare ability. New techniques in genetic sequencing are allowing them to explore birdsong neurons in unprecedented detail and find out which genes are responsible for the brain networks that enable finches to vocalize.

By studying the organization and genes of songbird brains, scientists like Colquitt can uncover the fundamental building blocks behind how different animals communicate. Doing so can inform the study and treatment of human speech disorders and deepen our understanding of how the brain is organized. “We can really understand at a very basic level—at a molecular level—how a brain is able to build itself to learn this incredibly complex behavior.”

Researchers say that while the gap between speech and birdsong is wide, applying genetics to bird brains can help us understand how the rare trait of vocal learning evolved and elucidate the mechanisms fundamental to an animal’s ability to communicate.

Bird Brains

In Colquitt’s lab, a dozen or so tan eggs sit warmed by the heat of an incubator. In the eggshell of each, a sizable hole reveals a heart no bigger than a pea. It beats with a barely perceptible fervor, pumping blood throughout a chicken embryo. But these embryos are unique—Colquitt has used viruses to deliver songbird genes into their brains.

Colquitt points to them with a flashlight, illuminating a crimson network of veins. “Avian models are great because you can do this,” he says, shifting the flashlight between embryos. “You can open up their eggs, you can watch them develop, you can put genes in their brain.”

He and his team are injecting each embryo with a set of genes they’ve identified as key in the development of finch brains. They’re investigating whether doing so could change the brains of chickens to be closer to those of songbirds. Through pioneering experiments like these, Colquitt is uncovering what makes songbird brains tick from a genetic perspective.

“Anytime you want to figure out how a circuit board works, you really want to know what the parts list of that circuit board is,” says NYU neuroscientist Michael Long, who describes Colquitt as a “rising star” in the field of birdsong neuroscience. “What Brad has done is created this beautiful kind of technicolor parts list.”

Colquitt’s research, he says, adds clarity to decades of work investigating songbird brains. Scientists have looked to birdsong as a model for vocal learning in humans since at least the 1990s, when neuroscientist Allison Doupe and speech scientist Patricia Kuhl published a formative paper outlining the similarities between the two.

“That particular review was a way of legitimizing this research in the eyes of a broader swath of people interested in speech,” says Colquitt. Since then, researchers have been exploring the brains of songbirds, hummingbirds, and parrots—all birds that learn their vocalizations—to understand the key networks that underlie vocal learning.

But these similarities are only half of why researching birdsong is so valuable. The other half lies in the way bird brains are organized. “[Humans] are 100 billion neurons; these guys are on the order of hundreds of millions of neurons. So it’s a smaller problem, and in the areas that we really care about, it’s much more tractable,” says Long.

While regions of the human brain devoted to speech are interconnected in a complex array of neural networks, the brains of songbirds have a specific region—called the song system—devoted exclusively to vocal learning. This allows researchers to focus on and ask fundamental questions about how vocal learning arises in the brain.

Illustration Credit: Aly Colasanti

Musical Genes

Back in his lab, Colquitt approaches a small box attached to the side of a birdcage and opens it. An apprehensive finch perks up, tilting its tan head to analyze him. Nestled warm under its rump, naked and wriggling, is a nestling. Its eyes haven’t opened yet, but the neurons in its brain have already started priming themselves to learn how to sing.

Some of the nestlings in the lab will be raised normally, hearing the songs of adult males to learn and develop their own vocalizations. Others, however, will be raised by females in isolation from any singing males. Comparing the two, Colquitt says, can help zero in on what genes are most important for vocal learning.

Both baby birds and humans depend on hearing vocalization early in life to develop correctly. During a critical period, when babies babble and nestlings subsing (the bird version of babbling), they are highly sensitive to the vocalizations of adults. Their brains depend on the sounds it hears from adult birds to encode an accurate neural model of how to speak and sing.

“If you take a baby bird and raise it in isolation from adult song, it’ll develop something called an isolate song,” Colquitt explains. Without the early exposure, the song systems of birds cannot correctly imitate the songs of their parents and will instead produce random vocalizations.

Comparing neurons between birds raised with song and birds raised without it can clue the researchers in to which cell types and genes are most important for the process of vocal learning. “Then we can go and test it,” Colquitt says. Once his team lands on a candidate gene, they can try modifying its expression in songbird brains to see how it impacts the process of vocal learning.

At his desk, Colquitt points to a map of cell types he’s made for the neurons of a Bengalese finch. It’s the “technicolor parts list” Long alluded to, and it’s impressive.

The figure shows colorful clusters made up of thousands of dots, each representing data from an individual neuron, or brain cell. The neurons are grouped based on similarities in their gene expression, with each cluster assigned a different color to denote a specific cell type.

The map was made using single-cell RNA sequencing, a technology that allows scientists to measure gene expression within individual cells. It has enabled neuroscientists to be far more precise in determining which neurons are expressing which genes in the brain.

In a typical experiment, researchers isolate individual cells from brain tissue. Each cell’s RNA molecules—which reflect which genes are active—are tagged with a unique molecular barcode. These barcodes make it possible to trace every RNA molecule back to its original cell, revealing which genes are expressed in which cells.

After the cells are broken open, their RNA is sequenced. Scientists can then determine which genes were expressed in each individual cell and at what relative levels. By comparing gene expression patterns across thousands of cells, researchers cluster them into distinct types and investigate which genes are important for the function of neurons in the brain’s song region.

By analyzing specific cells and their gene expression in birds that learned song versus those that did not, Colquitt can identify which neurons and genes are activated during vocal learning. “You can really start to figure out the nuts and bolts of what the important neurons are, and how they are changing during this process of learning.”

Approaching Aphasia

As a speech pathologist for Mayo Clinic, Rene Utianski has a unique appreciation for birds. “From the social linguistic perspective of how birds socialize with one another in different ways, and how that relates to social human communication, I think there’s a lot of value there.”

She studies disturbances in human speech like apraxia and aphasia, which cause individuals to lose their ability to communicate effectively by impairing their motor coordination. In the complexity of the human brain, Utianski says, diagnosing the cause of these conditions can be challenging because language regions are so interconnected with the rest of the brain. “There’s lots of different ways things can go wrong.”

What we do have, she says, is a strong understanding of the motor systems involved in speech. In particular, humans depend on an area of the brain called the basal ganglia to produce fine motor movements. “When I talk with patients, I will often describe it as the brake system,” says Utianski.

The basal ganglia are a group of interconnected brain regions deep within the brain. They carefully control which signals are allowed to come through and move the muscles animals use to vocalize. It’s key in motor learning, releasing dopamine to positively reinforce correct vocalizations.

When the basal ganglia aren’t working in the human brain, disorders affecting movement and speech occur. Understanding the specific mechanisms of speech disorder in humans is particularly challenging, Utianski says, because “there are just a lot more pieces and parts that are coordinating to make human speech.”

This is where understanding the brains of songbirds can be most impactful, says University of Texas Southwestern neuroscientist Todd Roberts. “We can get answers to things that would be very difficult to study in other species, because other species just simply don’t engage in this type of complex imitated behavior.”

Birds and humans share many similarities in their ability to learn, produce, and refine their vocalizations. In addition, the basal ganglia of birds function in ways that closely parallel those of humans, regulating the fine details of how a song is performed and adjusted over time.

However, the song circuitry in a bird’s brain—including its basal ganglia—is exclusively dedicated to learning and maintaining vocalizations. In humans, by contrast, language and speech systems are far more interconnected with other cognitive and motor networks.

With fewer component parts and less complex connectivity, researchers can focus on this specialized circuitry to understand the fundamentals of vocal production. “We can ask very specific questions,” Roberts says, like how does a gene that’s been implicated in autism selectively disrupt the networks of neurons and pathways in the brain.

His sentiments are echoed by both Long and Colquitt, who hope a mechanistic understanding of vocalization will inform human speech pathology. “We get this kind of real, full understanding between genes and behavior in a way that’s extremely difficult to be able to study for humans,” says Colquitt.

While Utianski hasn’t yet seen the research reach the human side of the issue, she hopes future collaborations between birdsong researchers and human speech pathologists can inform both fields. “I wholeheartedly believe in the power of translational science,” she says.

The Wonder of Birds

For Colquitt, the brains of songbirds aren’t just fascinating, they’re embedded with evidence of evolution’s ability to problem solve. “The problem being, how do you exist in a complicated world and adapt to a complicated world with complex social structures,” he says.

Back in the redwoods of Henry Cowell, birders quietly take in the conversations of their avian neighbors. Rich songs filled with information ring out into the forest, each one a reminder that human conversations aren’t the only ones of value.

Wilde, who’s been birding since he was a kid, doesn’t need to know what they mean to appreciate the songs. “I can’t tell who’s saying what to whom,” says Wilde, “but there’s a lot going on here.”

© 2017 Example Author / UC Santa Cruz Science Communication Program